Simulate age-year-space covariance using SPDE approach

Returns a function to use inside sim_distribution to generate the error term.

sim_ays_covar_spde(
  sd = 2.8,
  range = 300,
  model = "barrier",
  phi_age = 0.5,
  phi_year = 0.9,
  group_ages = 5:20,
  group_years = NULL,
  mesh,
  barrier.triangles
)

Arguments

sd

Variance (can be age specific)

range

Decorrelation range

model

String indicating "barrier" or "spde" to generate Q with

phi_age

Defines autocorrelation through ages. Can be one value or a vector of the same length as ages.

phi_year

Defines autocorrelation through years. Can be one value or a vector of the same length as years.

group_ages

Make space-age-year variance equal across these ages

group_years

Make space-age-year variance equal across these years

mesh

The mesh used to generate the precision matrix

barrier.triangles

the set of triangles in the barrier of the mesh for the barrier model

Value

Returns a function for use in sim_distribution.

Examples

##SPDE Approach

# \donttest{

## Make a grid
my_grid <- make_grid(res = c(10,10))

## Make a mesh based off it

my_mesh <- make_mesh(my_grid)
sim <- sim_abundance(ages = 1:10, years = 1:10) %>%
         sim_distribution(grid = my_grid,
                          ays_covar = sim_ays_covar_spde(phi_age = 0.8,
                                                         phi_year = 0.1,
                                                         model = "spde",
                                                         mesh = my_mesh),
                          depth_par = sim_parabola(mu = 200,
                                                   sigma = 50))
plot_distribution(sim,ages = 1:5, years = 1:5, type = "heatmap")


## Barrier Approach
sim <- sim_abundance(ages = 1:10, years = 1:10) %>%
         sim_distribution(grid = survey_grid,
                          ays_covar = sim_ays_covar_spde(phi_age = 0.8,
                                                         phi_year = 0.1,
                                                         model = "barrier",
                                                         mesh = survey_lite_mesh$mesh,
                                                         barrier.triangles =
                                                         survey_lite_mesh$barrier_tri),
                          depth_par = sim_parabola())
#> Warning: Arbitrary prior values chosen automatically. This may suffice for a first attempt, 
#>             but should be changed in any serious analysis.
plot_distribution(sim, ages = 1:5, years = 1:5, type = "heatmap")


# }